Visual Analytics: Data, Analytical and Reasoning Provenance
Analysts and decision makers are increasingly overloaded with vast amounts of data/information which are often dynamic, complex, disparate, conflicting, incomplete and, at times, uncertain. Furthermore, problems and tasks that require their attention can
- PDF / 265,500 Bytes
- 10 Pages / 439.36 x 666.15 pts Page_size
- 11 Downloads / 307 Views
Abstract Analysts and decision makers are increasingly overloaded with vast amounts of data/information which are often dynamic, complex, disparate, conflicting, incomplete and, at times, uncertain. Furthermore, problems and tasks that require their attention can be ambiguous, i.e. they are ill-defined. In order to make sense of complex data and situations and make informed decisions, they utilize their intuition, knowledge and experience. Provenance is fundamental for the user to capture and exploit effectively the explicit data and implicit knowledge within the decision making process. Provenance can usefully be considered at three conceptual levels, namely: data (what), analytical (how) and reasoning (why). This paper explores visual analytics in the exploitation of provenance within the decision making process. Keywords Analytical provenance • Data provenance • Hypothesis • Reasoning provenance • Visual analytics • Visualization
1 Introduction Analysts and decision makers are increasingly overloaded with vast amounts of data/information which are often dynamic, complex, disparate, conflicting, incomplete and, at times, uncertain. Furthermore, problems and tasks that require their attention can be ambiguous, i.e. they are ill-defined. In order to make sense of complex data and situations and make informed decisions, decision makers rely on explicit information and their implicit intuition, knowledge and experience. Moreover, to have confidence in a decision making process, it is necessary for them
M. Varga Seetru Ltd, Bristol, UK, e-mail: [email protected]; [email protected] University of Oxford, Oxford, UK C. Varga Seetru Ltd, Bristol, UK © Springer International Publishing Switzerland 2016 V.L. Lemieux (ed.), Building Trust in Information, Springer Proceedings in Business and Economics, DOI 10.1007/978-3-319-40226-0_9
141
142
M. Varga and C. Varga
to understand the sources of information and thus the value and trust that can be placed on every aspect of the process; i.e. the provenance [1–4]. Provenance is fundamental for the user to capture and exploit effectively the explicit data and implicit knowledge within a decision making process. Provenance can usefully be considered at three conceptual levels, namely: data, analysis and reasoning [5]. In essence it comprises the what (data), how (it was analyzed) and why (reasoning). This paper explores the application of visual analytics as an effective means of analyzing and understanding provenance in the explicit representation of the analytical and reasoning processes: how and why the data is used.
2 Data, Analytical and Reasoning Provenance There are three categories of provenance that play a role in visual analytics, namely: data provenance, analytical provenance and reasoning provenance. In order to understand findings/discoveries it is necessary to document the entire analysis process and retain all three types of provenance. Capturing the reasoning processes is by far the most challenging. • Data provenance considers the sourc
Data Loading...
